This article was downloaded by: [University of Sydney] On: 09 November 2011, At: 16:29 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Civil Wars Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/fciv20 The Dynamics of Warfare in Civil War Adam Lockyer a a University of Sydney E-mail: [email protected] Available online: 28 Jun 2010

To cite this article: Adam Lockyer (2010): The Dynamics of Warfare in Civil War, Civil Wars, 12:1-2, 91-116 To link to this article: http://dx.doi.org/10.1080/13698249.2010.484906

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. The Dynamics of Warfare in Civil War

ADAM LOCKYER

This article explains theoretically, and demonstrates empirically, the instrumental role of the balance of capabilities in shaping the form of warfare that develops in civil wars. It contends that the current common practice of labelling civil wars as either ‘guerrilla’ or ‘conventional’ (which is usually meant to accurately characterise the type of warfare throughout an entire civil war) is unable to fully encapsulate the dynamic nature of warfare in civil war. It is instead argued that the form of warfare frequently varies significantly across time and space in a single conflict. This article is divided into three sections. Section one examines recent advances in the understanding of warfare in civil wars. It identifies three categories of warfare in civil war: conventional, guerrilla and irregular. Section two builds on previous studies to develop the concept of the balance of capabilities. Finally, the article illustrates these theoretical insights through a discussion of the American and Somali Civil Wars.

There has been a recent upsurge in interest in the nature of warfare in civil war.1 Stimulated by the seminal work of Stathis Kalyvas, this fresh line of research is currently grappling with the implications of three typologies of warfare in civil war: conventional, guerrilla, and irregular (or ‘symmetrical non-conventional’).2 One such implication, which has yet to be properly explored, is that warfare in civil war is clearly more diverse than previous studies have acknowledged. Previously, warfare in civil wars has usually simply been labelled as ‘conventional’ or ‘guerrilla’; such labels being intended to accurately describe the type of warfare which took place throughout the conflict. However, accepting a triple typography Downloaded by [University of Sydney] at 16:29 09 November 2011 throws this practice into question, as this typography forces the dynamic nature of warfare in civil war into clearer view. By relying on this typography, we are able to observe that at different times, different forms of warfare dominated the fighting in, for example, the Chinese, South Vietnamese and Afghan civil wars. In addition, different forms of warfare have been fought simultaneously in different geographical regions of a single civil war. During the late-1980s in Angola, for instance, conventional warfare dominated the fighting in the south, while guerrilla warfare was prolific in the north. Indeed, it is now becoming clear that the significant longitudinal and cross-sectional variation in warfare is one of the defining characteristics of civil wars. This article argues that the warfare in civil war is highly dynamic. Indeed, civil wars frequently exhibit considerable spatial and temporal variation in their form of

Civil Wars, Vol.12, Nos.1–2 (March–June 2010), pp.91–116 ISSN 1369-8249 print/ISSN 1743-968X online DOI: 10.1080/13698249.2010.484906 q 2010 Taylor & Francis 92 CIVIL WARS

warfare. This prominent characteristic of civil wars has yet to be properly studied. The article extends beyond this argument by attempting to identify the causes of the variation in warfare in civil wars. In doing so, it builds upon the intuition of numerous civil war theorists by regarding the strategic balance between belligerents as the instrumental determinant of many phenomena in civil wars, including the course and nature of warfare.3 It makes several important theoretical advances on previous attempts at a balance of capabilities approach. First, whereas strategic balances have previously been understood and applied as a simple relative measure comparing belligerents’ capabilities, in this article, a model of the balance of capabilities is developed that includes both absolute and relative measurements. This innovation accounts for a belligerent’s selection of a strategy being a function of its own physical ability and its strength vis-a`-vis its opponent(s). Second, unlike previous attempts, the explanation of the balance of capabilities presented in this article allows for the possibility of there being more than a single insurgent actor. Finally, the variables are more sensitive and fluid than previous models. As will be demonstrated, this allows for increased dynamism, a stronger predictive element and a more sophisticated method of analysis overall. This article is divided into three sections. Section one develops a revised model of the balance of capabilities and explains its instrumental role in determining the form of warfare that comes to dominate a particular place and time in a civil war. Section two offers an improved means through which to measure both the balance of capabilities (the independent variable) and warfare (the dependent variable). In particular, this section makes an important contribution by submitting that the offensive and defensive characteristics of each form of warfare represent the best method of empirical identification of warfare in civil war. The third section demonstrates the role of the balance of capabilities in influencing the form of warfare that develops at a particular place and time, by empirically analysing the causes of the changes in the warfare in the American and Somali civil wars. The American and Somali civil wars are commonly pointed to in the literature as representing ideal types of conventional and irregular warfare, respectively. Even so, as this brief historical analysis demonstrates, there was significant spatial and temporal variation in even these cases. The article concludes that it was the balance Downloaded by [University of Sydney] at 16:29 09 November 2011 of capabilities that was responsible for this variation in warfare.

WARFARE AND THE BALANCE OF CAPABILITIES A common and recurring misnomer in the literature is the use of the terms ‘war’ and ‘warfare’ synonymously. In fact, however, the two terms are homonyms rather than synonyms. War refers to the political and philosophical nature of the conflict. As such, the type of war normally denotes the political issues at the heart of the conflict; for example, a ‘revolutionary war’, ‘anti-colonial war’ or ‘ethnic war’. Warfare, in contrast, signals the means through which the war is fought. Warfare is thus a function of the strategies employed by the contending belligerents. That is, different types of strategies, in different combinations, will produce alternate forms of THE DYNAMICS OF WARFARE IN CIVIL WAR 93

warfare. Hence observers, such as M.L.R. Smith, are correct when they argue that regardless of whether a civil war is fought through conventional or guerrilla strategies, the fundamental nature of war remains the same.4 This line of thought, however, fails to adequately acknowledge that the nature of the warfare would be radically different in a civil war in which the belligerents each employed conventional strategies from one in which the insurgent utilised a guerrilla strategy. There has been significant variation in the warfare that has characterised civil wars. To a much greater extent than interstate conflicts, warfare in civil war is dynamic, changing considerably over time and space. At a broad, idealised level, there are three forms of warfare in civil war: conventional, guerrilla and irregular. The first form of warfare, conventional, derives from the traditional image of warfare between states. Its guiding principle is to deploy the maximum amount of force at a decisive point and thereby destroy an opponent’s ability to resist in a ‘set piece battle’. In contrast, guerrilla warfare is characterised by the incumbent and insurgent actors employing different strategies. The insurgent actor attempts to avoid a direct test of strength, opting to structure its units into small, decentralised bands and striking the incumbent’s units and economic installations before fading back into the terrain and civilian population. In turn, the incumbent generally employs a strategy that attempts to isolate the insurgent’s forces from the civilian population, while protecting its own infrastructure. The principles of irregular strategy fundamentally differ from those of guerrilla and conventional strategies and, unlike them, are not well-established in the theoretical literature on civil war. It has long been assumed by historians that warfare in civil wars cannot always be divided into a simple conventional/guerrilla dichotomy. Historians writing on various civil wars have rejected the dichotomy in favour of triple strategic categories, for instance, by dividing the Confederate Army into conventional, guerrilla and militia units. Irregular warfare has been commonplace since the Middle Ages, when European lords organised civilians into local militias for their personal use and to repel marauding mercenaries and raiders.5 The modern-day equivalents of these earlier wars have been seen in Bosnia, Georgia, Haiti, Lebanon, Liberia, Sierra Leone, Somalia, Sri Lanka, the Sudan and the Democratic Republic of the Congo.6 Thus, historians and other observers have Downloaded by [University of Sydney] at 16:29 09 November 2011 been right to elevate irregular warfare to a prominence equal to that of conventional and guerrilla warfare. In political science, however, irregular warfare has frequently been misdiagnosed. The end of the Cold War and the subsequent outbreak of wars in failed states led many to report the appearance of a new type of civil war. These new civil wars went by a variety of names: ‘postmodern wars’,7 ‘wars of a third kind’,8 and ‘new wars’.9 Kalyvas correctly identified, however, that these analysts were not observing the emergence of a new type of war, but rather increasing instances of irregular warfare.10 Irregular warfare emerges in cases where the belligerents’ forces are too weak for conventional warfare, but are evenly matched, thus making guerrilla unnecessary. The only major difference between my conception and Kalyvas’ is in terminology. I use the term ‘irregular’ to describe the form of warfare 94 CIVIL WARS

that Kalyvas labelled ‘symmetrical non-conventional’ (SNC). My rationale for this change is twofold. First, although SNC has the virtue of being descriptive, it also suffers from carrying the connotation of being a ‘new’ term for a ‘new’ form of warfare. As the fundamental characteristics of irregular warfare can be traced back through history, I judged a new term to be inappropriate and misleading. Second, while SNC accurately describes the warfare, non-conventional warfare fails to precisely capture the strategy employed by each belligerent. That is, non- conventional strategy has already been largely appropriated by students of nuclear and non-conventional weapons. Therefore, I argue that ‘irregular’ – being sufficiently vague and ill-defined in the literature to allow redefinition – is a more suitable label. The definitions of conventional, guerrilla and irregular warfare are further developed in section two of this article. At this point, several questions become pertinent to this discussion. First, what causes the great spatial and temporal variation in warfare in civil war? Further, how can we predict which form of warfare will emerge at a particular place and time in a civil war? The remainder of this section will attempt to answer these questions.

The Balance of Capabilities As briefly discussed, warfare is the product of the clash of the belligerents’ individual strategies. Belligerents, in turn, adopt strategies subject to constraints imposed by their physical resources and those of their opponent. Only to a lesser extent will warfare be influenced by political objectives, geography and military culture.11 For example, the causes, location, and population involved in the Burmese, South Vietnamese, and Algerian civil wars were vastly different and yet, for the most part, they were all characterised by guerrilla warfare. This section will show that the belligerents’ method of war-making will principally be influenced by the distribution of military capabilities within the civil war system – that is, the configuration of the balance of capabilities. Within the civil war system I assume that there are only two types of actors: the incumbent and one or more insurgents. I take the incumbent to be the actor that controls the capital city of the civil war state.12 Capital cities are generally the centre Downloaded by [University of Sydney] at 16:29 09 November 2011 of bureaucratic, military, and political control and as such are highly-prized by most belligerents whose war aim is to win control over the entire state. Therefore, a belligerent will usually remain the incumbent until it loses control over the capital city.13 Indeed, even if an actor is effectively besieged, having no direct control over anything further away than the outskirts of the capital, it will remain the incumbent.14 Consistent with this definition, an incumbent may become an insurgent by losing control of the capital city, but staying in the game by retreating to rural areas to continue the fight.15 An insurgent is defined as a significant contender for control over part, or all, of the state’s territory, and who possesses the military means to pursue its goals. The organisation, capabilities and strategic objectives of insurgent groups certainly vary widely, yet this definition clearly distinguishes insurgents from other opposition groups such as non-violent dissidents and criminals.16 THE DYNAMICS OF WARFARE IN CIVIL WAR 95

Military Capabilities and Conventional, Guerrilla and Irregular Warfare The three forms of warfare in civil wars are the product of different configurations of the balance of capabilities. Conventional warfare is the result of a strong force confronting another strong force; guerrilla warfare is the outcome of a strong force opposing a relatively weak force; and irregular warfare is the product of a weak force challenging another weak force. There are two prerequisites for conventional warfare to appear in civil war. First, each belligerent must be able to field sufficient numbers of disciplined troops and equip them with military equipment. In practice, this normally means that both belligerents must have access to the full range of instruments commonly associated with a state: territory, manpower, a considerable material resource base, a bureaucracy and a relatively sophisticated military machine. In other words, conventional warfare is the result of opposing belligerents each having access to large volumes of military capabilities. The second prerequisite for conventional warfare to occur is that both protagonists must calculate that there is a good probability that they will be victorious by pursuing such a strategy. If the success of a conventional strategy is considered unlikely, a belligerent may consider dissolving its conventional units and converting to a guerrilla strategy. As a result, conventional warfare is the result of both absolute and relative variables. Guerrilla warfare is the product of an asymmetry in belligerents’ relative military capabilities. That is, in choosing to pursue a guerrilla strategy, a belligerent is acknowledging that it perceives its military capabilities to be inherently inferior to those of its opponent. However, belligerents that adopt a guerrilla strategy may not be weak in absolute terms, only when compared to their opponent. Hence, the size of guerrilla forces and their capabilities may vary greatly between cases. For example, when Mao pursued a guerrilla strategy, his army consisted of a sizeable 10,000 to 30,000 fighters. Nevertheless, Mao’s forces were opposed by armies three to 10 times greater in size, ensuring that a conventional confrontation with the enemy would have been disastrous.17 Finally, irregular warfare emerges when the belligerents remain weak in absolute

Downloaded by [University of Sydney] at 16:29 09 November 2011 terms, but have relatively balanced military capabilities. As such, this form of warfare will occur when the belligerents do not possess the capabilities to employ a conventional strategy and when guerrilla warfare is rendered unnecessary by the fact that the incumbent is not substantially stronger than the insurgent. Irregular warfare will therefore generally emerge where the state’s ‘structure, authority, power, law, and civil order have fragmented’.18 As the incumbent’s relative resources decline, its ability to project influence out from the capital also decreases. This process creates an opportunity for regional centres of authority to emerge. David Francis, for example, observed that ‘militias mushroom in weak, failed and collapsed states, where the state does not have control or monopoly of the threat of force’.19 The levelling of the belligerents’ capabilities leads to comparable methods of warfare. The collapse of the state frequently not only sees the government shift from relying 96 CIVIL WARS

on regular forces to militia and paramilitary forces, but also the emergence of warlords employing similar forces.

The Balance of Capabilities: Average Military Capabilities and Distribution At this point it is now possible to begin constructing an explanation of the balance of capabilities. A belligerent’s freedom to choose a strategy is restricted and shaped by two factors. The first is an absolute measure – the total value of the belligerent’s own capabilities. That is, the range of strategies available to a belligerent is limited by what it is physically able to achieve. The second factor influencing a belligerent’s strategic decisions is a relative measure – specifically, the measure of a belligerent’s relative strength vis-a`-vis its opponent(s). The first determinant of the form of warfare that occurs in a civil war is the average military capabilities. This variable is an absolute measure and is equal to the sum of the incumbent’s and the insurgent’s (or insurgents’) military capabilities, divided by the total number of belligerents. A belligerent’s military capabilities shape its own behaviour by permitting and restricting different strategies. Still, the overall form of warfare that takes hold in a civil war will be a confluence of all the belligerents’ strategies. Thus, warfare will be a function of the sum of all the belligerents’ military capabilities and the number of belligerents involved in the civil war. As a rule, the greater the number of belligerents in a civil war, the weaker each side is likely to be. The number of belligerents in a civil war will have a major influence on the military capabilities it is possible for each side to possess, as the number of belligerents roughly corresponds with the number of divisions of territory, fighters, and resources. In June 1990, for instance, the Liberian Civil War was seemingly reaching its climax with the National Patriotic Front of Liberia (NPFL) reaching the outskirts of Monrovia. Although, the NPFL numbered some 10,000 fighters, only several hundred core fighters were trained and relatively- disciplined, and with victory imminent, many of these fighters decided to break away and form the Independent Patriotic Front of Liberia (IPFL). The short-term consequences of the insurgent actor’s split into two mutually antagonistic belligerents was to weaken the NPFL, carve Monrovia into three different spheres of Downloaded by [University of Sydney] at 16:29 09 November 2011 control, and allow the Liberian government to stave off immediate defeat.20 To capture the importance that the number of belligerents has on the balance of capabilities, the first variable that is identified here is the average military capabilities, which is equal to the sum of all capabilities in the system divided by the number of belligerents. When applied to the balance of capabilities, the average military capabilities will range between a low and a high measure. The second variable that affects the balance of capabilities is the distribution of military capabilities in the civil war. It is nonsensical to assume that there will ever be a perfectly equal distribution of military capabilities between the belligerents, as the average military capabilities variable would suggest if presented alone. Hence, the division of capabilities must also be considered. In fact, unlike the concept of the ‘balance of power’ in international relations theory which is assumed to tend THE DYNAMICS OF WARFARE IN CIVIL WAR 97

towards equilibrium, the logic of the balance of capabilities suggests that the ‘natural state’ of the distribution of military capabilities will be in the incumbent’s favour. There are three reasons for this proposition. First, the incumbent is, by definition, partially in control of the coercive instruments of the state, and is therefore in a better position to extract resources. Even in a weak or failing state, the military capabilities at the disposal of the incumbent will generally be – however marginally – greater than those of its challengers. Second, even during peacetime, most states continue to prepare for the possibility of war, while most sections of society do not. As such, the incumbent will usually be better organised for war than the insurgent, at least at the outset of hostilities. Finally, if the balance of capabilities shifts in favour of an insurgent, one of three outcomes will usually occur. An insurgent with a greater distribution of military capabilities may (1) win the war, (2) capture the capital (driving the incumbent into rural areas), or (3) fractionalise and dampen its strategic advantage vis-a`-vis the incumbent. In the last situation, the division of the insurgent will result in a corresponding division of its military capabilities and leave the distribution of military capabilities in favour of the incumbent. When incorporated into the balance of capabilities, the distribution of military capabilities is a purely relative measure that will range between disparity and parity between belligerents. By considering the average military capabilities and distribution of military capabilities in light of the different characteristics of conventional, guerrilla and irregular warfare, it is possible to formulate some hypotheses about the form of warfare that will emerge at a particular time and place in a civil war. First, conventional warfare will be adopted when the average military capabilities are high and the belligerents mutually calculate that there is relative parity in the distribution of military capabilities. Second, guerrilla warfare will occur with high or low average military capabilities, when the distribution of military capabilities is disparate. Although the average military capabilities will influence the volume of firepower each belligerent can apply, it will not affect the fundamental nature of the warfare which occurs. Finally, irregular warfare will take place when the average military capabilities are low and there is parity in the distribution of military capabilities. In this case, although there is military symmetry in the civil war, neither Downloaded by [University of Sydney] at 16:29 09 November 2011 side possesses sufficient military capabilities to attain the high force-to-space ratio required for conventional offensives or to secure strong positional defences. Figure 1 illustrates how the two variables interact to influence the form of warfare that emerges in a civil war. However, this depiction remains static. Hence, two further factors must be taken into consideration in order to better capture the dynamics of warfare in civil war. First, longitudinal measurements of the balance of capabilities at different temporal points will generally show some variation in one or both variables. That is, the military capabilities of the belligerents will continually be in flux as soldiers are recruited or killed, territory changes hands, and military equipment is destroyed, purchased or captured. Foreign assistance can also quickly or gradually increase the capabilities available to the belligerents. Second, there may be spatial variation in 98 CIVIL WARS

FIGURE 1 THE BALANCE OF CAPABILITIES AND WARFARE IN CIVIL WARS

Disparity

Guerrilla and counter-guerrilla warfare

Distribution of military capabilities

Irregular warfare Conventional warfare Parity Low High Average military capabilities

the balance of capabilities. Certain geographic factors can potentially create regional variation in the balance of capabilities and therefore warfare. For instance, even belligerents with large militaries are unlikely to spread their forces evenly across an entire civil war state. Therefore, some areas will have higher concentrations of troops than others. Regional differences in terrain and internal political boundaries can also affect the particular balance of capabilities in a region.21 Accordingly, the balance of capabilities will constantly fluctuate. These sudden or gradual shifts in the balance of capabilities, across time and between regions, may cause the belligerents to change their type of strategy. There are two factors influencing a belligerent’s change in strategy: physical and cognitive. The value of a belligerent’s military capabilities imposes physical limitations on strategy. For instance, a belligerent that lacks a sufficient number of skilled and disciplined troops

Downloaded by [University of Sydney] at 16:29 09 November 2011 and conventional military assets (such as tanks, artillery and other heavy weapons) may well find it physically impossible to employ a conventional strategy. After this physical threshold in a belligerent’s military capabilities has been crossed, and it becomes materially able to change forms of strategies, there is no guarantee that a belligerent will instantly decide to do so. A belligerent will have to decide if the time is right to adopt a different strategy, and so the second determinant of whether a belligerent’s strategy changes is a construct of their perception of the balance of capabilities. As Robert McColl asserted, there ‘is no quantitative measure as to when this change must occur. It is a purely subjective matter that depends upon the objective realities of any given geographic and political situation’.22 That is, the second factor influencing the transition of strategy from one form to another will rest upon the calculations of the politico-military leadership of the contending THE DYNAMICS OF WARFARE IN CIVIL WAR 99

belligerents. As such, the exact point in which belligerents decide to change strategies will tend to differ between cases. This last point suggests that although the balance of capabilities is instrumental in shaping the behaviour of belligerents in civil war, it is not deterministic. At times, belligerents will miscalculate the balance of capabilities and independently choose an inappropriate strategy. For instance, a belligerent may choose to adopt a conventional strategy when it is still too weak and will subsequently be punished for its mistake. Such cases are littered throughout history. In the Greek Civil War, for example, the Communists’ abandonment of ‘guerrilla tactics led to the inevitable: the victory of the government army, which had superiority in numbers and fire- power’.23 Similarly, in 1946, the Viet Minh converted to a conventional strategy and was inflicted with heavy losses by the superior French forces as a result. In response, the Viet Minh quickly ‘renounced open combat, and dissolved its divisions and its regiments’ and reverted back to a guerrilla strategy.24

MEASURING THE BALANCE OF CAPABILITIES AND WARFARE IN CIVIL WAR The relationship between the balance of capabilities and warfare designates the former as the independent variable and the latter as the dependent. This section submits two novels means to measure both the balance of capabilities and warfare.

The Measurement of the Balance of Capabilities The balance of capabilities is deduced though an assessment of the belligerents’ military capabilities. That is, each belligerent’s military capabilities must first be ascertained before they can be combined in a judgement of the balance of capabilities. As a concept, military capabilities encapsulates a belligerent’s size, technology and fortitude in a single variable. Hence, in measuring a belligerent’s military capabilities for this article, I have employed a composite qualitative approach which is primarily comprised of three factors: troop strength, technology and contemporary reports. I argue that, when combined, these measures provide a Downloaded by [University of Sydney] at 16:29 09 November 2011 reasonably reliable judgement of each belligerent’s military capabilities and, more importantly, the state of each belligerent’s military capabilities relative to its opponent(s). Some may argue that other factors are central to calculating a belligerent’s military capabilities, such as organisational structure and training. However, both organisational structure and training contain an endogeneity problem for the balance of capabilities. That is, these factors not only make substantial contributions to the balance of capabilities but are also outcomes of a belligerent’s choice of strategy. Hence, organisational structure and training feed into both the independent and dependent variables. Troop strength is a popular measure of military capabilities.25 The main advantage of employing troop strength as a component of the measurement of military 100 CIVIL WARS

capabilities flows from it being common to all belligerents, easily comparable and relatively precise. Clearly, however, troop strength is deficient on its own, as superior technology and fighting ability can all potentially compensate for a larger force. The type and quantity of a belligerent’s military technology – that is, its weapons and equipment – is an integral component of that belligerent’s military capabilities. It is important to include military technology in the measurement of military capabilities because it can potentially make a significant contribution, dramatically increasing a belligerent’s capacity for offensive and defensive operations. Like organisational structure and training, however, incorporating military assets into the measurement of military capabilities represents possible endogeneity. The weapons a belligerent possesses may not only shape its chosen strategy, but the belligerent may also select weapons based upon its strategy. Nevertheless, I assume that in most instances, a belligerent will attempt to accumulate as much destructive potential as possible. Thus, the type, quantity, quality and sophistication of the military technology the belligerents employ will tend to be more of an expression of their military capabilities than their choice of strategies. Finally, a measurement of belligerents’ military capabilities should rely on historical accounts. I will therefore draw upon the observations of reporters, participants and other assorted sources to gauge the absolute and relative strength of the belligerents. Contemporary accounts can make an important contribution to accurately judging belligerents’ military capabilities. Furthermore, the importance of contemporary accounts lies in their ability to confirm the impact that troop strength and technology have on belligerents’ military capabilities. Indeed, the contribution of troops and technology are likely to vary significantly. For example, not all soldiers will be equally disciplined and willing to fight, and military technology will vary in its state of repair and appropriateness to the terrain. Thus, overall, marshalling the views of contemporary sources captures many of the intangible elements which troop numbers and technology neglect, while also engaging contemporary observers’ own evaluation of the balance of capabilities.

Measuring Warfare Downloaded by [University of Sydney] at 16:29 09 November 2011 Previous attempts at defining conventional, guerrilla and irregular warfare have been insufficient. Generally, definitions have simply pointed to historical cases (suggesting that ‘we will know it when we see it’26), looked for technological markers (i.e. tanks as a sign of conventional warfare), or looked at the relative strength of the contending sides. The first method clearly lacks rigor; the second fails to acknowledge that technologies can be employed in novel ways; and the final method obviously possesses an endogeneity problem for this particular study. As such, a fresh approach at empirical observation warfare is required. It is submitted here that the forms of warfare are differentiated from each other by their empirical characteristics. In other words, each form exhibits different operational features. This can be observed through a comparison of each type’s offensive and defensive operational elements. THE DYNAMICS OF WARFARE IN CIVIL WAR 101

Conventional Warfare Operationally, conventional warfare is characterised by a clear distinction between offensive and defensive actions. In offence, conventional strategy generally involves massive coordinated ‘pushes’ or ‘thrusts’ into enemy held territory. The key aim is to decisively engage and destroy the opposition’s forces through attrition, blitzkrieg or ‘normal progress’. It requires a belligerent to decrease the force-to-space ratio by concentrating its forces at the most advantageous location before advancing into enemy-held territory. In defence, conventional warfare assumes frontlines as the defining feature. The principal aim of fortified positions is ‘to repel attack, to protect people and property, to hold territory, and to minimize damage by the attacker.’27 Although not every inch of territory may be protected, strategically-important locations are prepared to be held against the enemy’s assaults for as long as possible. This emphasis on territory, in both defence and offence, is alluded to in the common reference to conventional warfare as ‘positional warfare’.28

Guerrilla and Counter-guerrilla Warfare In guerrilla warfare, the incumbent and insurgent apply very different strategies. This makes it unique among the types of warfare in civil war. The insurgent adopts a guerrilla strategy which involves it waging a protracted war of attrition while avoiding open engagements with the adversary. Offence in guerrilla strategy attempts to hold the initiative by engaging in surprise attacks, raids, sabotage and ambushes, thereby avoiding a pitched battle.29 An insurgent applying a guerrilla strategy endeavours to create the impression that its forces may strike anywhere and anytime, which compels the incumbent to spread its forces thinly in order to protect all important infrastructure. As a consequence, the incumbent has fewer available forces with which to aggressively pursue the guerrilla force. In defence, guerrilla forces do not attempt to defend territory or population. Instead, and in contrast to conventional strategy, guerrillas attempt to decrease the force-to-space ratio. As Che Guevara emphasised, ‘the essential task of the guerrilla fighter is to keep himself from being destroyed’.30 Defensive guerrilla strategy thus requires the insurgent to withdraw from territory when challenged by the enemy forces, even if this means Downloaded by [University of Sydney] at 16:29 09 November 2011 leaving highly-valued assets like family homes and sites of symbolic value unprotected. For this reason, Clausewitz appropriately likened the strategy to a cloud that parts as solid pressure moves towards it.31 In order to avoid an incumbent’s military forces, an insurgent who has adopted a guerrilla strategy will frequently attempt to blend into particularly rugged territory or a section of society. However, as a French colonel in Algeria identified, ‘this total dependence upon terrain and population is ... [also] the guerrilla’s weak point.’32 Thus, offensive counter-guerrilla strategies adopted by incumbents may attempt to isolate the guerrillas from their social support base and deny them readmission, tempt or force them out of cover and into a direct confrontation, or annihilate the social base from which the insurgency originates.33 Although the methods used in counter-guerrilla strategy are unusual within a conventional context, their primary 102 CIVIL WARS

strategic objective is similar: to locate the enemy and destroy it through superior manoeuvre and firepower. In defence, counter-guerrilla strategy requires the protection of important political, economic and military positions. However, this task can be problematic, as the insurgent, by striking at unpredictable places and at times, can create an ‘air defence’ dilemma for the incumbent. When constructing an air defence strategy, the required number of air defence units is more a function of the targets that must be defended than the number of enemy bombers. Therefore, in counter-guerrilla strategy, an incumbent with many potential targets will have to invest heavily in defending these installations by constructing elaborate conventional defences (such as trenches, bunkers and observation posts), regardless of the actual number of enemy guerrilla units.

Irregular Warfare Irregular offensives exhibit different characteristics in urban and rural areas. In urban areas, irregular offensives frequently resemble severe gang violence, with seemingly ‘confused street battles’34 and, in more recent times, the increasing use of civilian cars, motorbikes and trucks for swift hit-and-run attacks. In rural areas, unlike the thrusts of conventional warfare or the ‘hit and run’ strikes of guerrilla strategy, irregular offensive operations frequently resemble trickles. Irregular advances often follow the route of least resistance. In Sierra Leone, for example, the Revolutionary United Front (RUF)’s decentralised advance edged forward following the path of least resistance. A farmer from the Pujehun District in Sierra Leone recounted that: ‘[t]he RUF boys had two-way walkie-talkies. News of the easy advance was passed on. They said, “come, there’s no resistance. The [All Peoples’ Congress] is nothing.” Another group entered Pujehum [district] at Suluma, hearing the word.’35 Clearly, instead of concentrating its forces for a conventional drive, or using guerrilla infiltration, the RUF captured territory through a third form of offensive strategy, that is, capturing territory and asserting its control over population by advancing into unprotected space. In short, in offensive operations, irregular strategy attempts to achieve two objectives: to capture territory Downloaded by [University of Sydney] at 16:29 09 November 2011 and avoid battles of annihilation. In defence, belligerents employ an irregular strategy attempt to hold and defend territory and population. Indeed, militias, paramilitaries and youth gangs are frequently raised and paid for by belligerents to deny territory and population to the enemy. In practice, this characteristic is evidenced in a number of ways. One example is the roadblocks and checkpoints that frequently litter civil wars that are fought using irregular warfare. Across cases of civil war, the sturdiness of roadblocks varies dramatically, from concrete guard houses through to ‘a few stones or empty plastic crates’, as reported in Rwanda.36 Regardless of construction, the role of roadblocks remains to assert a belligerent’s control over a territory and population. The appearance of roadblocks and checkpoints in a civil war frequently create unmistakable frontlines. However, unlike the frontlines in conventional THE DYNAMICS OF WARFARE IN CIVIL WAR 103

warfare, frontlines in irregular warfare are relatively lightly-defended and frequently appear in large numbers. Indeed, a single city may be divided between several belligerents, each asserting its control over different suburbs; such was the case in Beirut, Mogadishu, Baghdad and Monrovia. In Monrovia, the different militias ‘raced around in pickup trucks on their side of the city, clutching assault rifles, rocket launchers and two prized anti-aircraft guns’.37 Several characteristics of irregular warfare appear similar to conventional or guerrilla warfare. Therefore, some may question the extent to which empirical features distinguish irregular warfare from conventional or guerrilla warfare. However, as an amalgamation, the characteristics of irregular warfare justify its status as a third distinct form of warfare. For instance, although raids and ambushes are a common feature of both guerrilla and irregular strategies, in a sharp deviation from the central tenet of guerrilla/counter-guerrilla warfare, only in irregular warfare do both the incumbent and insurgent forces employ similar techniques. Likewise, roadblocks are commonly used to control territory and population in both counter-guerrilla and irregular strategy. However, as these small obstacles represent easy targets for powerful opponents, it is generally only in irregular warfare that insurgent actors also use these constructions. Table 1 summarises the defining operational characteristics of conventional, guerrilla and irregular strategies. Nonetheless, although it is analytically useful to discuss the three different types of warfare in civil war separately, they are not mutually exclusive. As the table suggests, there is frequently substantial blurring between the idealised analytical categories. Moreover, as this article argues, it is common for regular, guerrilla and militia units to be found operating in different regions, or at different times, in a single civil war. These different types of units, applying different techniques, often ascribe civil war with a seemingly confusing complexion. The following section attempts to take steps towards overcoming some of this confusion by identifying the variables that influence which form of warfare develops at a particular place and time.

Downloaded by [University of Sydney] at 16:29 09 November 2011 TABLE 1 OPERATIONAL AND ORGANISATIONAL CHARACTERISTICS OF THE DIFFERENT FORMS OF STRATEGY IN CIVIL WAR

Conventional Strategy Irregular Strategy Guerrilla Strategy

Units Regular soldiers Militias and Guerrillas paramilitaries Logistics Most incorporated Mixed Most incorporated into the state into society Organisational structure Centralised Decentralised Decentralised Force-to-space ratio High Medium Low Offensive characteristics Concentrated thrusts Dispersed trickles, Strikes, sabotage, raids raids Defensive characteristics Strong fixed Weak fixed Evasion 104 CIVIL WARS

SPATIAL AND TEMPORAL VARIATION IN THE BALANCE OF CAPABILITIES AND WARFARE The purpose of this section is to empirically illustrate the fluid nature of the balance of capabilities and the impact this feature has on warfare in civil war. It briefly reflects on the relationship between the balance of capabilities and warfare in two conflicts: the American and Somali civil wars. The American and Somali civil wars are, respectively, popular examples of conventional and irregular warfare. However, as the American and Somali civil wars illustrate, even in these extreme examples, there was significant spatial and temporal variation in warfare.

Spatial Variation and the American Civil War The American Civil War is frequently touted as a perfect example of a civil war fought by conventional warfare. There are good reasons for this practice. In the Eastern Theatre, Union and Confederate generals employed what General Schofield labelled a ‘territorial strategy’.38 The focus was on capturing the enemy’s territory or, in the South’s case, invading the North so as to draw the Union army away from Virginia. What is lesser known, however, is that outside the narrow corridors of the Eastern Theatre, warfare assumed a very different character. In fact, guerrilla warfare characterised much of the remaining fighting in Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Missouri, Tennessee, Texas and north- western Virginia.39 In these regions, small bands of loyalist or secessionist guerrillas harassed the occupying armies. This section will demonstrate the relationship between regional variation in the balance of capabilities and the warfare that develops in different locations by comparing the conventional warfare of the Eastern Theatre with the warfare in Missouri.

Conventional Warfare and the Eastern Theatre The popular image of the American Civil War is of two grand armies, marching out onto an open field in rank and file before commencing to slaughter each other with musket, cannon and sabre; in other words, a civil war fought by conventional Downloaded by [University of Sydney] at 16:29 09 November 2011 warfare. This view is not erroneous, as most of the strategically-significant warfare was conventional in nature. This was particularly true of the Eastern Theatre, where conventional strategy was available to belligerents due to their significant and relatively-even strength in the zone. The balance of capabilities in the Eastern Theatre was characterised by high average military capabilities and parity in the distribution of military capabilities. Although the Union possessed an overall quantitative superiority in troop strength, the Confederacy was successful in deploying a comparable number of soldiers to all major battles.40 For instance, at the First Battle of Bull Run on 21 July 1861, the Union fielded 28,452 men compared to the Confederacy’s 32,232. At the Battle of Gettysburg, the North had 88,289 to the South’s 75,000.41 Furthermore, in the Eastern Theatre, both belligerents issued their forces with weapons of similar sophistication THE DYNAMICS OF WARFARE IN CIVIL WAR 105

and quality and thus neither side possessed ‘an asymmetrical advantage’ in tech- nology.42 With roughly equal numbers facing off on the battlefield, both belligerents had a reasonable expectation that they could be victorious by pursuing a conventional strategy. The Eastern Theatre was fought through a combination of conventional forms of offensive operations, namely manoeuvre and attrition. The respective armies attempted to manoeuvre their forces into a favourable position and engage the enemy in a decisive battle. This was implied by General Grant’s 1864 statement that: ‘[m]y general plan now was to concentrate all the force available against the Confederate armies in the field ... I [then] arranged for a simultaneous movement all along the line’.43 The two sides threw tens of thousands of soldiers into the pursuit of a decisive victory. In defence, the belligerents’ forces frequently sought shelter behind a ‘protective wall or entrenchment, to shield themselves from the attackers’ rifles’.44 However, due to the introduction of rifled artillery and small arms, the American Civil War saw the first signs that the nineteenth century fortresses were becoming obsolete. During the civil war, traditional structures were gradually replaced by other forms of defensive technologies, namely ‘redoubts’ and ‘rifle pits’.45

Guerrilla Warfare in Missouri The American Civil War is not normally associated with guerrilla warfare. However, guerrilla warfare was in fact the principal form of warfare across much of the country. Michael Fellman noted that in Missouri, when the Confederate’s regular army was absent, the ‘war often assumed a deadly guerrilla nature as local citizens took up arms spontaneously against their neighbours. This was a war of stealth and raid, without a front, without formal organization, with almost no division between the civilian and the warrior.’46 As Fellman suggested, it was the presence of large numbers of troops and militiamen on the side of the Union, and the absence of any significant body of troops on the side of the Confederates, that transformed the warfare in Missouri into guerrilla. The balance of capabilities and warfare in Missouri was typical of much of the country outside that occupied by both a large Union and Confederate regular army. Downloaded by [University of Sydney] at 16:29 09 November 2011 In Missouri, the Union had between 40,000 and 60,000 regular and part-time soldiers scattered throughout the major towns and garrisons.47 In comparison, at any one time, there was unlikely to have been more than 300 active Confederate guerrillas, known as bushwhackers.48 Generally, these guerrillas were organised into bands of between 20 and 60 men. Clearly, the Union possessed a vast advantage in military capabilities. The guerrillas’ only advantage was in weaponry. For the most part, Union soldiers were armed with single-shot muzzle-loading carbines and sabres, whereas the guerrillas employed revolving pistols and repeating rifles, which allowed the guerrilla bands to lay down ‘a tremendous volume of fire’.49 However, the guerrillas’ weaponry only marginally improved the bushwhackers’ military capabilities and therefore the overall distribution of military capabilities was highly disparate. 106 CIVIL WARS

The Missouri guerrillas had a number of favourite methods for attacking Union infrastructure and troops. The most militarily significant were sabotage operations against Missouri’s railway network. It was reported that, due to railroad sabotage, over a hundred miles of railroad had been destroyed across northern Missouri alone and St Louis was practically cut off from the rest of the state.50 Another popular offensive technique was ambush. As historian Albert Castel documented: [the guerrillas’] favorite device was to lie in ambush besides a road along which some unwary Union patrol was approaching. When the soldiers drew close, the guerrillas would suddenly gallop out of the bush, screeching the rebel yell and blazing away with their pistols. In a matter of seconds the affair would be over. The partisans would strip the dead, round up their horses, and disappear once more into the thicket...51 As this description suggested, the Confederate bushwhackers did not attempt to hold and defend territory. Instead, after conducting sabotage, raid or ambush operations, they would disperse into the civilian population or terrain or cross the border into Confederate territory. Throughout the civil war, Union forces in Missouri employed a counter-guerrilla strategy. In offence, the Union forces used a combination of eradication and isolation techniques. The most notorious of these was an isolation strategy called ‘General Order No. 11’. When in June 1863, Brigadier General Thomas Ewing assumed command of the Union forces of the District of the Border, he estimated that two-thirds of the civilian population in the district had kin riding with the guerrillas and wrote: ‘I can see no prospect of an early and complete end to the war on the border, without a great increase of troops, so long as the families remain here.’52 As there were no more troops to deploy, Ewing issued General Order No. 11, which stipulated that the Missouri border counties be depopulated of civilians by 9 September 1863.53 During the civil war, coercive isolation strategies, like those employed in Missouri, were popular counter-guerrilla techniques of both Union and Confederate authorities. In defence, the Union forces were restricted to defending fixed positions against guerrilla raids. As Bowen observed, ‘the Union forces were rendered relatively immobile by the necessity to protect fixed points such as urban areas, supply depots, and lines of communication.’54 Downloaded by [University of Sydney] at 16:29 09 November 2011 Therefore, although the Union forces enjoyed a vast numerical superiority, a substantial proportion of these soldiers were tied down guarding towns and other strategic sites. Spatial variation in the strategies that belligerents employ is a frequent feature of civil war. Clearly, even the American Civil War experienced significant spatial variation in warfare. Distance, topography and political borders act to separate territories, which create independent regional balances of capabilities and thus different forms of warfare.

Temporal Variation and the Somali Civil War If the American Civil War is normally classified as an example of conventional warfare, the Somali Civil War has frequently been cited as an ideal example of THE DYNAMICS OF WARFARE IN CIVIL WAR 107

irregular warfare. However, as with the American Civil War, the Somali Civil War was, in fact, also characterised by forms of warfare other than that by which it is typically identified. Unlike the American Civil War, the most prominent variation in warfare in Somalia was temporal rather than spatial. That is, the conflict passed through a number of different phases, each characterised by a different form of warfare. This section will briefly illustrate the relationship between temporal changes in the balance of capabilities and warfare by sketching the course of the Somali conflict from 1982 until the multinational intervention in 1992. Over the course of these 10 years, the warfare moved from guerrilla to irregular warfare following changes in the balance of capabilities. This section will specifically outline the changing nature of the conflict between the Somali Armed Forces (SAF) and the insurgent Somali National Movement (SNM).

Guerrilla Warfare Phase (1982–90) Following the disastrous 1977–78 Ogaden War, opposition to the presidency of Siyaad Barre grew across many sections of Somali society. Hostility toward Siyaad was particularly fierce in northern Somalia, which felt that an insufficient amount of state investment had flowed to the north. In 1982, the SNM, drawing its support principally from the Isaaq clan-family lineage, commenced guerrilla operations against the SAF. At the outbreak of hostilities, the average military capabilities in Somalia was quite high. The Siyaad presidency had been characterised by the militarisation of Somali society. In 1976, Somalia was the only country with a per capita annual income of less than $150 which was classified as a major arms importer.55 By the early 1980s, it was likely that there were more small arms in Somalia than people. Nevertheless, the distribution of military capabilities remained firmly in the SAF’s favour. In 1982, the SAF had 120,000 soldiers,56 while the SNM may have constituted some 2,000–3,000 guerrillas.57 In addition, although the SAF had been mauled during the 1977 war with Ethiopia, it continued to possess a large qualitative and quantitative advantage in weaponry over the SNM. In 1982, the SAF could field some 150 tanks, 200 armoured fighting vehicles and small numbers of combat aircraft.58 In Downloaded by [University of Sydney] at 16:29 09 November 2011 contrast, the SNM was armed mainly with light weapons. Hence, overall, the average military capabilities in Somalia was reasonably high and the distribution was disparate. As the theory expects, this configuration of the balance of capabilities resulted in the SNM adopting a guerrilla and the SAF a counter-guerrilla strategy. The SAF’s counter-guerrilla tactics targeted the general Isaaq population, which was seen as providing ‘an inexhaustible source of guerrilla recruits for the SNM’.59 Siyaad decided to separate the guerrillas from their social base by forcing Isaaq civilians over the border into Ethiopia. The SAF destroyed watering wells, ruined grazing grounds and indiscriminately raped and killed Isaaq civilians.60 The strategy was both brutal and effective. Almost a million Isaaqs fled to Ethiopia and large tracts of northern Somalia were depopulated.61 A foreign traveller reported that: ‘one can ride for half a day hardly seeing a soul except for some shepherds, all of whom 108 CIVIL WARS

carry guns’.62 To defend larger towns and garrisons, the SAF deployed artillery, armoured vehicles and aircraft. These proved devastating when the SNM concentrated its limited and lightly-armed guerrilla fighters to assault these targets. For instance, when the SNM launched a series of major offensives against the SAF in northern Somalia, between May and December 1988, the SAF responded by deploying tanks, artillery and aircraft to inflict thousands of causalities on the hapless guerrillas.63 The SNM’s guerrilla strategy emphasised raids, ambushes and evasion. The SNM guerrillas were armed with a mixed assortment of small arms, including old bolt-action rifles and captured assault rifles. Grenades and land-mines were also in plentiful supply. Yet, during the early 1980s, the SNM possessed no tanks, armoured vehicles or aircraft. The limited size and potency of the SNM meant that offensive operations were generally restricted to ambushing SAF patrols and raiding government positions. Indeed, according to one report, during this phase, ‘SNM offensives amounted to guerrilla attacks on military outposts and occasional publicity stunts, such as the kidnapping of 10 French aid workers from a refugee camp near the Ethiopian frontier in January 1987.’64 In defence, the SNM seemed to blend into the civilian population and northern Somalia’s rugged terrain. ‘The SNM is a close expression of the people’, one report noted, ‘and is as much at home in its environment as the ubiquitous camel.’65 Following a raid, the SNM also had the option of fleeing over the Ethiopian border where Addis Ababa allowed it, until 1988, to pass freely through the Isaaq refugee camps.

Irregular Warfare Phase (1990–92) During most of the 1980s, although the SAF was unable to defeat the SNM, it had effectively checked its expansion. However, in the late 1980s, spurred on by the example of the SNM, all the major Somali clans began to create their own forces. These belligerents encouraged their clansmen serving in the SAF to desert, taking with them all their weapons and equipment. At the close of 1990, the proliferation of clan-based militias left Siyaad in control of no more than 10 to 15 per cent of Somalia and, by 1991, Siyaad ‘was often referred to as the major of Mogadishu, Downloaded by [University of Sydney] at 16:29 09 November 2011 because that was the only area of the country he controlled. The military had disintegrated into what looked like clan militia.’66 On 27 January 1991, Siyaad plundered the Central Bank and fled south by tank to the capital of his native clan, Kismayu.67 Lewis reported that the: political geography of the Somali hinterland in 1992, consequently, closely resembled that reported by European explorers in the 19th century, spears replaced by Kalashnikovs and bazookas. These clan areas could only be entered or traversed by outsiders (people of other clans, foreigners), with the consent of the locals and, usually the payment of appropriate fees or ‘protection’.68 The balance of capabilities had undergone a dramatic change. First, although the total military capabilities in Somalia remained relatively steady (i.e. there had been THE DYNAMICS OF WARFARE IN CIVIL WAR 109

no substantial injection or removal of troops, weapons or equipment from the country), the average military capabilities decreased as those existing resources were divided between more belligerents. In 1991, there were at least 10 major belligerents contesting the civil war, along with countless smaller groups.69 The growth of these militias was generally at the expense of the SAF, whose members deserted to their respective clan militias. The disintegration of the SAF was also responsible for moving the distribution of military capabilities toward parity. By 1991, the SAF had effectively ceased to exist as a coherent organisation and its troops and equipment were divided up by the competing militias.70 Thus, while the average military capabilities moved towards lower levels, the distribution of military capabilities moved in the direction of parity. These changes in the balance of capabilities are depicted in Figure 2. As the SAF continued to dissolve, its organisational structure, procurement of resources, and tactics became ‘hardly distinguishable from the insurgent groups’.71 The 1991 Battle of Mogadishu ‘was basically a confrontation between Hawiye clan militias on one side, and Daarood elements of the Somali army and armed Daarood civilians on the other, rather than a battle between an insurgency and a state that no longer existed.’72 Indeed, of the estimated 15,000 combatants fighting in Mogadishu, less than 1,500 were said to be organised into units recognisable as being conventional forces. Over the period 1990–93, these ‘decentralised and barely institutionalised militias developed their own dynamic of violence’.73 That is, the characteristics of the belligerents’ strategies began to conform to that of irregular warfare. While light weapons remained the norm in most skirmishes, some of the rival militia groups also had access to small numbers of artillery, mortars and

FIGURE 2 THE BALANCE OF CAPABILITIES AND WARFARE IN THE SOMALI CIVIL WAR

Disparate Phase 1 Downloaded by [University of Sydney] at 16:29 09 November 2011

Distribution of military capabilities

Phase 2 Parity Low High Average military capabilities 110 CIVIL WARS

even tanks.74 However, the most popular offensive weapons of the irregular phase were converted Land Rover and Toyota four-wheel-drives. These vehicles (called ‘technicals’) were frequently equipped with a heavy-calibre machine gun, an anti-aircraft gun, or a recoilless rifle and could carry up to 14 militiamen who were usually armed with Soviet-style weapons.75 The militiamen in these vehicles ‘roamed the cities and roadways plundering, extorting, and killing’.76 The technicals of the rival militias raced through the streets of Mogadishu in a ‘constant cat-and-mouse game’.77 The United Nations chronicled the situation as follows: armed clashes tended to take the form of wild, chaotic exchanges of fire, featuring front-lines which could shift fifty or one hundred kilometers in a day as lines of defense disintegrated and regrouped. Supply lines were ad hoc to nonexistent, relying mostly on looting; occasionally a [technical] was captured or destroyed simply because it ran out of fuel.78 Although the belligerents were unable to mount large coordinated offensives, the near-constant street skirmishes within Mogadishu took a very heavy human toll; between 1991 and 1994, annual deaths from fighting in Somalia averaged nearly 55,000 people.79 In defence, the key features of the warfare were the ‘innumerable roadblocks’ which were guarded by young militiamen.80 The roadblocks were usually guarded by youths armed with machine guns and rocket propelled grenades.81 Roadblocks served two purposes. First, they acted to guard against rival clan militia infiltration. Second, they were a major source of revenue, levying large tolls from merchants and foreign travellers. The comparative evidence from the United States and Somalia seems to support the claim that warfare in civil war is highly dynamic. The main cause of this fluidity is the belligerents’ military capabilities, which are constantly in flux. Belligerents will not generally be able to maintain large quantities of troops and equipment in all areas, at all times. This will cause substantial spatial variation in the balance of capabilities. A comparison of the Eastern Theatre and Missouri during the American Civil War highlights this point. In addition, the course of a civil war is frequently defined by the respective belligerents’ military successes and failures and the Downloaded by [University of Sydney] at 16:29 09 November 2011 formation and termination of political alliances, all of which will impact the balance of capabilities and, in turn, the selection of warfare strategy. The course and changing nature of the Somali Civil War underlines this point.

CONCLUSION This article set out to achieve three objectives. First, it aimed to caution against the increasing popularity of coding civil wars as being fought by a single form of warfare. Due to the dynamic nature of warfare in civil wars, it will rarely be possible to code one civil war as entirely conventional, guerrilla, or irregular. Indeed, this article argued that attempts to do so overlook one of the distinguishing characteristics of warfare in civil war: its dynamism. On top of this argument, the THE DYNAMICS OF WARFARE IN CIVIL WAR 111

article sought to theoretically explain how the balance of capabilities is instrumental in influencing the form of warfare that develops at a particular place and time. Finally, through two brief historical analyses, the article empirical demonstrated the significant variation in warfare in civil war and the role that the balance of capabilities plays in causing this dynamism. The balance of capabilities approach allows conditional predictions to be made on the future course and nature of civil wars. How would current civil wars be affected should resources be added or withdrawn from the conflict? In Afghanistan, for instance, the balance of capabilities approach suggests that the withdrawal, or scaling back, of American forces will have major consequences for the way in which the conflict is fought. This is particularly true should the Afghan Army unsuccessfully expand its current capabilities. The withdrawal of foreign forces will shift the distribution of military capabilities significantly towards parity, while moving the average military capabilities towards lower levels. The result is likely to be irregular warfare between the Afghan government, the Taliban and perhaps additional insurgent actors. The possible return of warlord politics will have major consequences for how the international community approaches Afghanistan, and being forewarned of the likely affect withdrawal will allow appropriate policy preparations to be made.

ACKNOWLEDGEMENTS The author would like to thank Gil Merom, Anastasios Panagiotelis, Doreen Tan Fong Chen and the two anonymous reviewers for their astute comments on earlier versions this paper.

NOTES

1. Stathis N. Kalyvas, The Logic of Violence in Civil Wars (New York: Cambridge University Press 2006); Laia Balcells, Rivalry and Revenge: Making Sense of Violence Against Civilians in Conventional Civil Wars, HiCN Working Paper 51 (Brighton: Households in Conflict Network 2008). 2. Stathis N. Kalyvas, ‘Warfare in Civil Wars’ in Isabelle Duyvesteyn and Jan Angstrom (eds) Rethinking the Nature of War (Abingdon: Frank Cass 2005) pp.88–104. 3. See Lawrence Freedman (ed.) Military Intervention in European Conflicts (Oxford: Blackwell Downloaded by [University of Sydney] at 16:29 09 November 2011 Publishers 1994) p.9; James D. Fearon, ‘Why Do Some Civil Wars Last So Much Longer Than Others?’, Journal of Peace Research 41/3 (2004) pp.275–301; T. David Mason, Joseph P. Weingarten, Jr. and Patrick J. Fett, ‘Win, Lose, or Draw: Predicting the Outcome of Civil Wars’, Political Research Quarterly 52/2 (1999) pp.239–68; Paul Collier, Anke Hoeffler and Ma˚ns So¨derbom, ‘On the Duration of Civil War’, Journal of Peace Research 41/3 (2004) p.268; Patrick M. Regan, ‘Third-Party Interventions and the Duration of Intrastate Conflicts’, Journal of Conflict Resolution 46/1 (2002) pp.55–73. 4. For a clear example of confusing war with warfare, see Table 1 (p.21) in M.L.R. Smith, ‘Guerrillas in the Mist: Reassessing Strategy and Low Intensity Warfare’, Review of International Studies 29/1 (2003) pp.19–37. 5. Mark Charles Fissel, The Bishops’ Wars: Charles I’s Campaigns Against Scotland, 1638–1640 (Cambridge: Cambridge University Press 1994) pp.174–95. 6. The warfare in Liberia, Sierra Leone and the Democratic Republic of Congo are discussed in William Reno, Warlord Politics and African States (London: Lynne Rienner 1998). 7. Mark Duffield, ‘Post-Modern Conflicts: Warlords, Post-Adjustment States and Private Protection’, Civil Wars 1/1 (1998) pp.65–102. 112 CIVIL WARS

8. Edward E. Rice, Wars of a Third Kind: Conflict in Underdeveloped Countries (Berkeley, CA: University of California Press 1988). 9. Mary Kaldor, New and Old Wars: Organised Violence in a Global Era (Cambridge: Polity Press 1999); Donald M. Snow, Uncivil Wars: International Security and the New Internal Conflicts (Boulder, CO: Lynne Rienner 1996); Mats Berdal, ‘How “New” are “New Wars”? Global Economic Change and the Study of Civil War’, Global Governance 9/4 (2003) pp.477–502. 10. Stathis N. Kalyvas, ‘“New” and “Old” Civil Wars: A Valid Distinction?’, World Politics 54/1 (2001) pp.99–118. 11. For a comprehensive list of all the elements that influence strategy, see, Williamson Murray, Alvin Bernstein and MacGregor Knox, The Making of Strategy: Rulers, States and War (Cambridge: Cambridge University Press 1996). 12. Also, see Marika Landau-Wells, Capital Cities in Civil Wars: The Locational Dimension of Sovereign Authority, Crisis States Occasional Paper 6 (London: London School of Economics 2008). My motivation for employing the criteria is purely analytical and, unlike Landau-Wells, I make no normative claim that domestic and international actors necessarily perceive the belligerent in control over the capital as the legitimate sovereign of the entire state. Although, in sympathy with Landau-Wells’ claim, I have found that historically there has been some correlation between the actor in possession of the capital and the side considered to be the ‘legitimate government’. In the 1834 Spanish Civil War, for example, although France, England and Portugal were supporting the Cristina government, in September 1837, they agreed ‘to recognize Carlos if he took Madrid, and by secret and joint instructions empowered their ambassadors in Paris to take this momentous step as soon as Carlos was in possession of his capital city’. Philip E. Mosely, ‘Intervention and Nonintervention in Spain, 1838–39’, Journal of Modern History 13/2 (1941) p.204. 13. Defining the incumbent in such a manner has several precedents in the empirical literature. Ngoga, for example, assumed that the capture of the Ugandan capital city instantly resulted in a transfer of governmental power. Pascal Ngoga, ‘Uganda: The National Resistance Army’ in Christopher Clapham (ed.), African Guerrillas (Oxford: James Currey 1998) pp.91–106. 14. The case of the Rabbani government in Afghanistan provides an excellent example of a ‘besieged government’. In addition, Maley argued that much of the fighting over Kabul during the Afghan Civil War was for its ‘symbolic’ importance. See chapter 8 and William Maley, The Afghanistan Wars (New York: Palgrave 2002) p.201. 15. This was exactly the case of the Former Government of Rwanda (FGOR) which after losing control of Kigali, withdrew to Zaire and began aggressive guerrilla operations against the new Rwandan government. For more detail see, Wm Cyrus Reed, ‘Guerrillas in the Midst’ in Christopher Clapham (ed.) African Guerrillas (Oxford: James Currey 1998) pp.134–54. 16. See Daniel L. Byman, Peter Chalk, Bruce Hoffman, William Rosenau and David Brannan, Trends in Outside Support for Insurgent Movements (Santa Monica: RAND 2001) pp.4–5; On the aims of terrorism and the methods see, Grant Wardlaw, Political Terrorism (New York: Cambridge University Press 1989), in particular see, pp.34–58; Bruce Hoffman, Inside Terrorism (New York:

Downloaded by [University of Sydney] at 16:29 09 November 2011 Columbia University Press 1999). Some scholars do not, however, view the differences between civil war and communal conflicts as being fundamentally different, but rather a similar phenomenon at a different level of violence. See, for example, Sarah Kenyon Lischer, ‘Causes of Communal Warfare: Fear and Feasibility’, Studies in Conflict and Terrorism 22/4 (1999) pp.331–55. 17. Andrew C. Janos, ‘Authority and Violence: The Political Framework of Internal War’ in Harry Eckstein (ed.) Internal War: Problems and Approaches (London: The Free Press 1964) p.644. 18. Alice Hills, ‘Warlords, Militias and Conflict in Contemporary Africa: A Re-examination of Terms’, Small Wars and Insurgencies 8/1 (1997) p.35; Duffield (note 7) pp.66–102. 19. David J. Francis (ed.) Civil Militias: Africa’s Intractable Security Menace? (Aldershot: Ashgate 2005) p.2. 20. Herbert Howe, ‘Lessons of Liberia: ECOMOG and Regional Peacekeeping’, International Security 21/3 (1996) pp.145–76; Stephen Ellis, ‘Liberia 1989–1994: A Study of Ethnic and Spiritual Violence’, African Affairs 94/375 (1995) pp.165–97. Conversely, the collapse of any belligerent would result in more open territory, available recruits and weapons for the remaining belligerents to integrate. Under these circumstances, the surviving belligerents’ military capabilities would be expected to increase, although, unevenly. THE DYNAMICS OF WARFARE IN CIVIL WAR 113

21. See, for instance, Noel C. Fisher, War at Every Door: Partisan Politics and Guerrilla Violence in East Tennessee, 1860–1869 (Chapel Hill: The University of North Carolina Press 1997) p.3. 22. Robert W. McColl, ‘The Insurgent State: Territorial Bases of Revolution’, Annals of the Association of American Geographers 59/4 (1969) p.622. 23. See Amikam Nachmani, ‘Civil War and Foreign Intervention in Greece: 1946–49’, Journal of Contemporary History 25/4 (1990) p.512. 24. Phillippe Devillers, Histoire du Viet-Nam de 1940 a` 1952 (Paris: Editions de Seuil 1952) p.166 in Nathan Leites and Charles Wolf, Rebellion and Authority: An Analytical Essay on Insurgent Conflicts (Chicago: Markham Publishing 1970) p.59. 25. See, for example, J. David Singer, Stuart Bremer and John Stuckey, ‘Capability Distribution, Uncertainty and Major Power War, 1820–1965’ in Bruce M. Russett (ed.) Peace, War, and Numbers (Beverly Hills, CA: Sage Publications 1972) pp.19–48; Frank W. Wayman, J. David Singer and Gary Goertz, ‘Capabilities, Allocations, and Success in Militarized Disputes and Wars, 1816–1976’, International Studies Quarterly 27/4 (1983) pp.497–515; Gary Goertz and Paul F. Diehl, ‘Measuring Military Allocations: A Comparison of Different Approaches’, The Journal of Conflict Resolution 30/3 (1986) pp.553–81. 26. Smith (note 4) p.21. 27. David Tarr, ‘Defense as Strategy: A Conceptual Analysis’ in Stephen. J. Cimbala (ed.) National Security Strategy: Choices and Limits (New York: Praeger 1983) pp.217–35. 28. Mao Tse-Tung, Selected Military Writings of Mao Tse-Tung (Peking: Foreign Languages Press 1963). 29. Andrew C. Janos, ‘Unconventional Warfare: Framework and Analysis’, World Politics 15/4 (1963) p.643. See also, V. K. Anand, Insurgency and Counter-Insurgency: A Study of Modern Guerrilla Warfare (New Delhi: Deep and Deep Publications 1985) pp.17–18. 30. Che Guevara, Guerrilla Warfare (London: Souvenir Press 1998) p.15. 31. See Carl von Clausewitz, On War, ed. and trans. by Michael Howard and Peter Paret (New York: Everyman’s Library 1993) p.580. 32. Roger Trinquier, Modern Warfare: A French View of Counterinsurgency, trans. Daniel Lee (New York: Praeger 1964) p.63. 33. Gil Merom, How Democracies Lose Small Wars (Cambridge: Cambridge University Press 2003); Gil Merom, ‘Strong Powers in Small Wars: The Unnoticed Foundations of Success’, Small Wars and Insurgencies 9/2 (1998) pp.38–63. 34. Michael Roddy, ‘Liberian Rebels Attack Central Monrovia’, Reuters 27 Jul. 1990. 35. David Keen, Conflict and Collusion in Sierra Leone (New York: Palgrave 2005) p.84. 36. Rome´o Dallaire, Shake Hands with the Devil: The Failure of Humanity in Rwanda (London: Arrow Books 2004) p.277. 37. Ellen Kinckmeyer, ‘Battle Lines Leave some Liberians locked in Hunger/Fighting Persists Outside Capital’, Associated Press 9 Aug. 2003. 38. Ulysses S. Grant III, ‘Military Strategy of the Civil War’, Military Affairs 22/1 (1958) p.21. 39. Fisher (note 21) p.3. 40. In 1863, the Union army numbered some 2,128,948 men compared to the Confederacy’s 1,082,119 Downloaded by [University of Sydney] at 16:29 09 November 2011 soldiers. See Micheal Clodfelter, Warfare and Armed Conflicts: A Statistical Reference to Casualty and Other Figures, 1500–2000 (Jefferson: McFarland and Co. 2002) p.305. 41. Ibid. pp.306–14. A study of some 60 battles discovered that, on average, the Union had a 37 per cent numerical advantage over the Confederacy. See Peter Maslowski, ‘To the Edge of Greatness: The United States, 1783–1865’ in Williamson Murray, Macgregor Knox and Alvin Bernstein (eds) The Making of Strategy: Rulers, States and War (Cambridge: Cambridge University Press 1994) p.236. 42. Mark Grimsley, ‘Surviving Military Revolution: The U.S. Civil War’ in MacGregor Knox and Williamson Murray (eds) The Dynamics of Military Revolution, 1300–2050 (Cambridge: Cambridge University Press 2001) p.76. 43. Grant (note 38) p.24. He later revised his stated plan to ‘fight it out on this line if it takes all summer’. See Grimsley (note 42) p.83. 44. Russell F. Weigley, A Great Civil War: A Military and Political History, 1961-1865 (Bloomington, IN: Indiana University Press 2000) p.33. 45. Redoubts were earthen defensive works, generally ‘rectangular and usually with two or more artillery emplacements or bations at the corners. A ditch or moat surrounded walls carefully sloped to provide 114 CIVIL WARS

maximum deflection of shots fired into them.’ John F. Bradbury, Fort No.5: A Civil War Field Fortification in Springfield, Greene County, Missouri, Unpublished manuscript: available through ‘State Historical Society of Missouri Civil War Collection’ (1984) p.10. Rifle pits consisted of a timber or stone reinforced trench and an earthen parapet. Rifle pits were frequently, but not exclusively, used in conjunction with redoubts. 46. Michael Fellman, Inside War: The Guerrilla Conflict in Missouri during the American Civil War (New York: Oxford University Press 1989) p.23. Also see Kenneth W. Noe, ‘Exterminating Savages: The Union Army and Mountain Guerrillas in Southern West Virginia, 1861–1862’, in Kenneth W. Noe and Shannon H. Wilson (eds) The Civil War in Appalachia: Collected Essays (Knoxville, TS: The University of Tennessee Press 1997) pp.104–30. 47. Dale E. Davis, Guerrilla Operations in the Civil War: Assessing Compound Warfare During Price’s Raid (Fort Leavenworth: General Staff College 2004) p.26. 48. Don R. Bowen, ‘Quantrill, James, Younger, et al.: Leadership in a Guerrilla Movement, Missouri, 1861–1865’, Military Affairs 41/1 (1977) p.42. Elsewhere, Bowen had rounded the number down to 200 guerrillas, see Don R. Bowen, ‘Guerrilla War in Western Missouri, 1862–1865: Historical Extensions of the Relative Deprivation Hypothesis’, Comparative Studies in Society and History 19/1 (1977) p.30. Davis had estimated that 450 guerrillas participated in the August 1863 raid on Lawrence, Kansas, which is commonly reported to have been the largest of its kind and involved every available Missouri guerrilla fighter. See, Davis (note 47) p.2. 49. Richard S. Brownlee, Gray Ghosts of the Confederacy: Guerrilla Warfare in the West, 1861–1865 (Richmond, VA: Louisiana State University Press 1958) p.24. 50. Ibid. p.24. 51. Albert Castel, ‘Quantrill’s Bushwackers: A Case Study in Partisan Warfare’ in John T. Hubbell (ed.) Battles Lost and Won: Essays from Civil War History (Westport, CN: Greenwood Press 1975) p.180. 52. The War of the Rebellion: A Compilation of the Official Records of the Union and Confederate Armies, Serial 1, Vol. 22, Part 2 (Washington, DC: Government Printing Office 1880–1901) p.428 as reproduced in Davis (note 47) p.23. 53. Albert Castel, ‘Order No. 11 and the Civil War on the Border’, Missouri Historical Review 57/4 (1963) pp.357–68; Thomas Goodrich, Black Flag: Guerrilla Warfare on the Western Border, 1861– 1865 (Bloomington, IN: Indiana University Press 1995) pp.96–105. General Order No. 11 was used in conjunction with the Union’s ‘assessment’ policy. Essentially, ‘assessments’ were aimed at forcing suspected Confederate sympathisers to financially compensate loyal Unionists for damages done by guerrillas. At first, the policy met with some success, however, it was open to fraud and other abuses, which, over time, diminished its effectiveness. See, W. Wayne Smith, ‘An Experiment in Counterinsurgency: The Assessment of Confederate Sympathizers in Missouri’, The Journal of Southern History 35/3 (1969) pp.361–80. 54. Bowen, ‘Guerrilla War in Western Missouri’ (note 48), p.30. 55. Hussein M. Adam, ‘Somalia: Militarism, Warlordism or Democracy?’ Review of African Political Economy 19/54 (1992) p.20. 56. By 1985, however, it is likely that the SAF numbered half this figure, with the International Institute of Strategic Studies and the Economist Intelligence Unit both placing the size of the SAF at 62,550. Downloaded by [University of Sydney] at 16:29 09 November 2011 See The Economist Intelligence Unit, Country Profile: Ethiopia, Somalia, Djibouti (London: EIU 1986) p.31. 57. Paul B. Rich, ‘Warlords, State Fragmentation and the Dilemma of Humanitarian Intervention’, Small Wars and Insurgencies 10/1 (1999) p.82. 58. IISS, The Military Balance, 1981–1982 (London: The International Institute for Strategic Studies 1981) p.64. 59. Said S. Samatar, Somalia: A Nation in Turmoil (London: The Minority Rights Group 1991) p.19. 60. Robert Gersony, ‘Why Somalis Flee: A Synthesis of Conflict Experience in Northern Somalia by Refugees, Displaced Persons and Others’, International Journal of Refugee Law 2/1 (1990) pp.4–55; Samatar (note 59) p.19; Anonymous, Somalia: A Long-Term Human Rights Crisis (London: Amnesty International 1988) p.38. 61. Gersony (note 60) p.5. 62. Gerald Prunier, ‘A Candid View of the Somali National Movement’, Horn of Africa 13–14/3–4/1–2 (1990) p.107. Furthermore, Ogaden refugees fleeing Ethiopian counter-guerrilla operations were encouraged to move to northern Somalia. As Lewis explained: ‘male Ogadeni refugees in northern THE DYNAMICS OF WARFARE IN CIVIL WAR 115

Somalia, who have long been subject to illegal recruitment into Somalia’s armed forces, have been conscripted as paramilitaries militias to fight the SNM and man checkpoints on the roads. Ogadeni refugees have been encouraged to take over the remains of Isaaq shops and houses in what are now ghost towns. Thus, those who were received as refugees guests have supplanted their Isaaq hosts, many of whom – in this bitterly ironic turn of fate – are now refugees in the Ogaden.’ Ioan M. Lewis, ‘The Ogaden and the Fragility of Somali Segmentary Nationalism’, African Affairs 88/353 (1989) p.577. 63. Samatar believed there to have been 5,000 SNM casualties in this offensive, see Samatar (note 59) p.19. Ethiopia had allowed the SNM use its territory and supplied some indirect military assistance to the guerrilla groups. According to one account, the ‘Ethiopian “presence” is everywhere obvious (in the form of tinned food, various objects and so on), but Ethiopian military aid seems relative limited.’ Prunier (note 62) p.115. 64. Robert Powell, ‘Somali Rebels Seek Credibility as Force Independent of Ethiopia’, Reuters 9 Jun. 1988. 65. Prunier (note 62) p.112. 66. Samuel M. Makinda, Seeking Peace from Chaos: Humanitarian Intervention in Somalia (Boulder, CO: Lynne Rienner 1993) p.17. 67. Gerard Prunier, ‘Somalia: Civil War, Intervention and Withdrawal, 1990–1995’, Refugee Survey Quarterly 15/1 (1996) p.45. Also see, Richard H. Shultz, ‘State Disintegration and Ethnic Conflict: A Framework for Analysis’, Annals of the American Academy of Political and Social Science 541 (1995) pp.75–88. 68. Ioan M. Lewis, Blood and Bone: The Call of Kinship in Somali Society (Lawrenceville, NJ: Red Sea Press 1994) p.231. 69. A (incomplete) list of belligerents in 1991 would include: the SNM, the Somali Democratic Movement (SDM), United Somali Congress (USC), United Somali Congress-Aydid (USC-Aydid), Somali People’s Movement (SPM), Somali Salvation Democratic Front (SSDF), Somali National Front (SNF) and Southern Somali National Movement (SSNM), United Somali Front (USF), Somali Democratic Alliance (SDA), Somali National Democratic Union (SNDU), Somali African Muki Organisation (SAMO), and Somali National Union (SNU). 70. SAF weapons and equipment that were not pilfered, had long since fallen into disrepair. The SAF’s aircraft, for instance, had ‘all been grounded, one by one, as a result of poor maintenance rather than SNM fire’. Prunier (note 62) p.108. Also see IISS, The Military Balance, 1991–1992 (London: The International Institute for Strategic Studies 1991) p.119. 71. Jutta Bakonyi and Kirsti Stuvøy, ‘Violence and Social Order Beyond the State: Somalia and Angola’, Review of African Political Economy 32/104 (2005) pp.365–66. Also see Belachew Gebrewold, ‘Civil Militias and Militarisation of Society in the Horn of Africa’ in David J. Francis (ed.) Civil Militia: Africa’s Intractable Security Menace? (Aldershot: Ashgate 2005) pp.187–212; Roland Marchal, ‘Forms of Violence and Ways to Control it in an Urban War Zone: The Mooryaan in Mogadishu’ in Hussien M. Adam and Richard Ford (eds) Mending Rips in the Sky (Lawrenceville: Red Sea Press 1997) pp.193–209. 72. Daniel Compagnon, ‘Somali Armed Movements: The Interplay of Political Entrepreneurship and

Downloaded by [University of Sydney] at 16:29 09 November 2011 Clan-Based Factions’ in Christopher Chapman (ed.) African Guerrillas (Oxford: James Curry 1998) p.77. 73. Bakonyi and Stuvøy (note 71) p.366; Ismail I. Ahmed and Reginald Herbold Green, ‘The Heritage of War and State Collapse in Somalia and Somaliland: Local-Level Effects, External Intervention and Reconstruction’, Third World Quarterly 20/1 (1999) pp.113–27. 74. See John Sislin and Fredric S. Pearson, Arms and Ethnic Conflict (Lanham: Rowman and Littlefield 2001) p.38. 75. Michael Hedge, ‘Drug Gives Young Gunmen Courage’, The Washington Post 9 Dec. 1992, p.2. 76. David D. Laitin, ‘Somalia: Civil War and International Intervention’ in Barbara Walter and Jack Snyder (eds) Civil War, Insecurity, and Intervention (New York: Columbia University Press 1999) p.148. 77. Donatella Lorch, ‘As U.N. Girds to Leave Somalia, Renewed Fighting’, The New York Times 27 Feb. 1995. 78. United Nations Development Office for Somalia (UNDOS), Study on Governance in Gedo Region (Nairobi: United Nations Development Programme (UNDP) 1998) p.75. 116 CIVIL WARS

79. This is an average over the five years of 4,500 per month. Patrick M. Regan, Civil Wars and Foreign Powers: Outside Interventions in Intrastate Conflict (Ann Arbor, MI: University of Michigan Press 2000) p.80. Bryden estimates that in fighting over Mogadishu alone in 1992 nearly 30,000 people were killed, see Matthiew Bryden, ‘Somalia: The Wages of Failure’, Current History 94/591 (1995) pp.146–47. 80. Bakonyi and Stuvøy (note 71) p.366. 81. Marshall V. Ecklund, ‘Task Force Ranger vs. Urban Somali Guerrillas in Mogadishu’, Small Wars and Insurgencies 15/3 (2004) p.56. Downloaded by [University of Sydney] at 16:29 09 November 2011